Monday, January 16, 2006

Questioning of Science Behind Helena Claim Appears as BMJ Rapid Response

My letter, which questions the validity of a claim that the observed 14% drop in heart attack deaths in New York City between 2003 and 2004 following its smoking ban (implemented in July 2003) was attributable to the smoking ban and therefore provides evidence to support the claim of a much larger heart attack decline in Helena was published today as on online Rapid Response to the original Helena article.

The letter makes four basic points:

1. The claim that the observed decline in heart attack deaths in New York City was attributable to the smoking ban is based on questionable science as it suffers from post hoc, ergo propter hoc reasoning. The simple fact that a decline in heart attack deaths followed the implementation of the smoking ban does not mean that the smoking ban was the cause. Similar reasoning could lead one to attribute as an effect of the same smoking ban the observed 9.1% increase in hypertensive heart disease deaths in the City.

2. Even if there were observed changes attributable to the ban, they are not as large as suggested, because heart attack deaths were already falling substantially in New York City prior to the implementation of the smoking ban. Based on pre-existing trends in heart attack deaths during the period 2000-2003, one would have expected a 9% decline in heart attack deaths in the City in 2004 anyway (in the absence of a smoking ban). It seems sloppy to attribute the entire 14% decline to the smoking ban, even without regard to the first problem (#1 above).

3. The data from New York City refer to deaths from heart attacks, not heart attack incidence as was reported for Helena and Pueblo. Because heart attack mortality is affected strongly by medical treatment, changes in mortality could reflect treatment improvements over time. In fact, the New York Health Department itself, in reporting these data, concluded that: "The reduction in cardiac deaths accounted for most of the overall decrease. The decline reflects a continuing improvement in medical care, including better control of blood pressure, cholesterol and management of patients with cardiac events."

The Rest of the Story

As Jacob Sullum correctly noted about me in highlighting these arguments on Reason Online's Hit & Run blog: "Michael Siegel, who supports smoking bans ... has been challenging his fellow anti-smoking activists to stop abusing science in service of their cause."

This is exactly right. My commentary on this issue has nothing to do with supporting or not supporting smoking bans. I support them. But it does have to do with the misuse of science to try to support the cause. Ultimately, my fear is that the credibility of tobacco control advocates will be hurt if they continue to use poor science to back up their public claims. Eventually, the public and policy makers will catch on - but they won't know when we are making legitimate claims and when we are not. Like the boy who cried wolf, we'll eventually get no response when we really need one.

No comments: